Avaliação de um método de rastreamento de canal baseado em filtros de Kalman para beamtracking

Abstract

Recent studies on Fifth-Generation Mobile communications systems (5G) point to massive MIMO systems, beamforming technologies and the millimeter waves as a way to deal with the large demand for data transmission from the current applications of teleinformatics systems. The 5G mobile communication is challenging due to the short length of mmWaves, which suffer greatly from path losses, atmospheric absorption and block attenuations, and narrow radiation beams from large arrays, which suffers from the imminence of misalignment due to the terminal mobility. The beamtracking is a set fundamental periodic processes to ensure beam alignment over time. However, it is of great importance that the beamtracking algorithms have low training complexity, so as not to affect the system throughput, and low processing complexity, so that it can be executed in mobile device’s wireless interfaces in order to reduce the feedbacks. Kalman (KF) filters can provide a powerful framework for estimation of channels with low complexitys for beamtracking applications. This master thesis aims to study of the use of Kalman filters as the base beamtracking algorithm for millimeter wave massive MIMO channel using linear and planar antenna arrays. Three types of KFs are implemented: the EKF, the UKF and the EnKF, as well as its deterministic version, the DEnKF. The KFs algorithms are evaluated in six case studies: in high and low SNR regimes, terminal mobility with high-level mobility, channels with varied number of clusters and scatterers, high-dimensional systems from large arrays, beamforming architecture limited in levels of phase quantization, and heuristics errors of predetermination of filter parameters. A qualitative analysis is performed based on least squares algorithms. Computational simulations were performed to evaluate the algorithms.